Optimal design of district metered areas based on improved particle swarm optimization method for water distribution systems

Abstract

Abstract Although partitioning of water distribution systems (WDSs) into district metered areas (DMAs) is challenging, it can be effectively used for refined management and leakage control. A two-step novel process for DMA partitioning is proposed in this study, i.e. clustering and dividing. The first step is to cluster nodes through an improved METIS graph partitioning method. The second step is to optimize the location of flowmeters and gate valves on boundary pipes by obtaining the feasible solutions. The good solutions that constitute the Pareto front were produced, which could be a tough and time-consuming task. The paper proposes the innovative and efficient dividing phase: (a) selecting the important boundary pipes by hydraulic analysis; (b) using the improved particle swarm optimization algorithm; (c) proposing three objective functions. The proposed method is applied to Modena and EXNET networks to demonstrate its feasibility.